What Is TextQL?

TextQL is a cloud AI analytics platform built around “Ana,” an AI data analyst that answers questions, builds dashboards, and, more recently, generates Data Apps. Ana's accuracy rests on the TextQL Ontology: a git-backed semantic model written in TQL, TextQL's own modeling language, that defines objects, metrics, and relationships over connected warehouses and apps. TextQL connects to a wide set of sources (Snowflake, Databricks, Oracle, SAP HANA, Dremio, Power BI and Fabric semantic models, and SaaS apps), offers an Azure single-tenant deployment, and prices on usage with a free tier.

Where TextQL Falls Short

• The ontology is hand-built in a proprietary language. Someone has to write and maintain TQL for every object and metric. Questions outside what has been modeled degrade sharply, and the model drifts as schemas change.
• Questions are bounded by the model. Because Ana answers through the ontology, TQL's expressiveness limits what can be asked. Novel, cross-domain, or long-tail questions are where users hit the wall.
• Cloud SaaS, not on-prem. TextQL runs as a hosted service (multi-tenant or Azure single-tenant). Data and metadata flow through TextQL's environment — a non-starter for institutions whose data cannot leave the perimeter.
• Usage-based cost. Pricing is consumption-based; costs rise as more people and agents ask questions, and multi-step analyst sessions consume more compute and tokens per question.
• Analyst workbench, not a search layer. TextQL is optimized for an analyst-style UI (dashboards, notebooks, apps). Enterprises increasingly need a search layer that agents — Copilot, ChatGPT, Claude, Teams — call directly for fast, deterministic answers.

How Tursio Is Different

• No modeling language to learn. Tursio infers the context graph — joins, dimensions, measures, aliases, PII — automatically from schema, query history, and BI definitions, then tunes it from feedback. The expensive semantic work is done once and shared by every user and agent.
• Open-ended questions. Tursio's operator-tree planner handles questions nobody modeled in advance, grounded to real database values.
• On-premises first. Single Docker image inside your environment with RBAC, audit, Entra ID, and BYO LLM keys.
• Agent-native search. Exposed over MCP and a Search API to every major agent surface; in production, ~90% of Tursio queries already come from agents.
• Flat pricing. Flat license, about 3¢ per query in tokens billed to your own LLM subscription.

Ontology vs Context Graph

TextQL and Tursio agree that AI over enterprise data needs a semantic layer. They disagree on who builds it. TextQL's ontology is authored by people in TQL and versioned in git — precise where it exists, absent where it does not. Tursio's context graph is inferred by the system from the data and its usage, then refined by people only where answers miss. The result is broad coverage on day one, and accuracy that improves with use instead of with modeling effort.

Side-by-Side Comparison

Bottom Line

TextQL is a capable AI analyst for teams willing to model their business in TQL and run in TextQL's cloud. Tursio removes the modeling project, answers questions beyond any predefined model, and runs inside your perimeter on a flat license — built for the agents that are now the primary consumers of enterprise data.